In silico analysis of regulatory regions of the hsp70 and hsp83 genes of different origin

The programs MEME and MatInspector (GENOMATIX Software) were used to align the HSPA1 regulatory sequences of camel with corresponding orthologous human genes. Therefore, major regulatory elements such as TATA boxes, HSEs, Sp1 and NF-Y-binding sites were determined in the promoters of the HSPA1A, HSPA1B and HSPA1L genes, which comprise a cluster linked to the major histocompatibility complex (MHC) locus in all mammalian species studied thus far. In the course of this analysis, we detected a 112 bp insertion in the hspA1L gene promoter of humans and other primates (e.g., Gorilla gorilla and chimpanzee Pan troglodytes) [19]. Intriguingly, this primate-specific sequence includes two homeodomain-binding sites. Despite the high homology of hsp70 regulatory regions observed in vertebrates, particularly mammalian species, there are a certain number of deletions, insertions and substitutions that discriminate the species hsp70 promoters and may in principle underline different promoter strengths and specificity even within mammalian species (S1 Fig.)

In contrast to the studied mammalian species, the Diptera species used in this study drastically differ in the arrangement of the hsp70 promoters while exhibiting a similar architecture for hsp83 gene promoters. In our previous study, we generated a lambda library from the genomic DNA of S. singularior and sequenced five S. singularior hsp70 genes designated S1—S5, which are located in tandem or are in an inverted orientation and comprise the Hsp70 gene cluster in this species [21]. All hsp70 genes in D. melanogaster contain highly homologous promoter regions, with four similarly spaced HSEs localized at the same distance from the TATA box [23]. On the other hand, the orthologous genes in S. singularior share only comparatively conserved coding regions and the first 50 bp upstream of the transcription start site, which includes a TATA box and the first HSE (see S2 Fig.). In contrast with known Drosophila species, in S. singularior hsp70 genes, the regions upstream of the transcription start sites and the first HSE and downstream of the stop codon (50 bp area) do not exhibit significant homology, and 5’-flanking regions contain different number and spacing of HSEs [21] (see also S2 Fig.). In addition to HSEs, the hsp70 regulatory regions (130 bp upstream the TATA box) in D. melanogaster and other Drosophila species usually contain several GAGA-sites, which are necessary for rapid transcription induction [24]. Functional GAGA sites in Drosophila are usually represented by GAGAG or GAGAGAG motifs or contain several GAG trinucleotides separated by spacers with an uneven number of nucleotides (one, three or five) [25]. The search performed for canonical GAF-binding sites using the MatInspector program failed to detect the presence of these sequences in S. singularior hsp70 promoters. In contrast to Drosophila species, orthologous S. singularior promoters contain only single GAG motifs. Furthermore, the TATA box in S. singularior hsp70 promoters is represented by a TATATA sequence in contrast to the TATAAA motif observed in Drosophila. The analysis failed to detect any additional known regulatory motifs in common within the promoters of all five S. singularior hsp70 genes. Interestingly, another pair of orthologous hsp genes (i.e., hsp83) present in both compared species (S. singularior and D. melanogaster) exhibits comparatively high homology and shares a similarly arranged single HSE found at nearly the same distance from the TATA box [22].

Camel hsp70 promoters normally function in human cells and exceed the strength of human promoters under steady conditions

It is known that in mammals, hsp70 family genes belonging to the HSPA1 cluster are linked with the MHC class III region and include three genes. Two of these genes (HSPA1A and HSPA1B) form a tandem repeat and are stress induced, while the third (HSPA1L), which is localized in an inverted orientation, is constitutively expressed (Fig. 1A) [19,26].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. A. General arrangement of the HSPA1 cluster in mammals. B. The level of transcription activity of individual hsp70 gene promoters comprising the cluster based on the measurement of transient luciferase luminescence driven by constructs carrying different human and camel hspA1 promoters. (i)—HSPA1A, (ii)—HSPA1B and (iii)—HSPA1L. The signal levels are shown as the ratio of the intensity of the luminescence of firefly (pFF) and renilla (pRL) luciferase. C. EMSA experiments with human and camel HSPA1L promoters fragments: 1—H. sapiens promoter without 112 bp fragment specific for primates only, 2—isolated H. sapiens 112 bp fragment, 3—C. dromedarius promoter. *** and ### p ≤ 0.001, ** p ≤ 0.01, * p ≤ 0.05.

https://doi.org/10.1371/journal.pone.0115536.g001

We developed sixteen constructs where a reporter gene (firefly luciferase) ORF was placed under the control of HSPA1 group promoters of different origin and architecture (see S3 Fig. for details). Specifically, to investigate the transcriptional activity of inducible HSPA1A promoters, we generated five constructs that included different regulatory regions from corresponding human and camel genes. The first two constructs included the camel HSPA1A regulatory regions of different lengths (-413…+31 bp and-413…+192 bp in relation to the transcription start site). Analogous constructs for the human HSPA1A gene comprised-505…+25 bp and-505…+221 bp. These constructs include two HSEs, a TATA box, and several motifs providing constitutive expression of mammalian hsp70 genes such as Sp1- and NF-Y-binding sites (S3 Fig.). Because our previous analysis revealed an additional potential start codon (ATG) in the camel HSPA1A 5`-UTR [19], in the fifth construct, we substituted it for TTG to examine whether such a mutation would somehow influence gene expression.

Luminescence measurements indicated that under normal physiological conditions in human HEK293 cells (similar results were obtained in human HDF cell culture, data not shown) transformed with the obtained constructs, the activity of the camel HSPA1A promoter significantly exceeded (two times) the activity of the orthologous human promoter (Fig. 1Bi). However, after temperature elevation, the expression level of both constructs was similar. The mutation of the additional upstream ATG within the camel HSPA1A 5`-UTR did not influence construct expression; hence, this site apparently has no functional significance.

Analysis of the promoter activity of the second inducible gene in the cluster (HSPA1B) was performed exploring two constructs of different lengths for the camel gene (-738…+8 bp and-738…+195 bp with respect to the transcription start site) and two analogous constructs for the human gene comprising-733/+35 bp and-733…+206 bp (S4 Fig.). Measurements of promoter strength in the luciferase assay failed to reveal significant differences in the compared constructs either under normal conditions or after heat shock (Fig. 1Bii). The described constructs contained two HSEs, and in general, heat shock led to a 5–6-fold increase in the expression of all of the HSPA1A and HSPA1B constructs regardless of the species of origin (the results of all construct activity measurements are summarized in Fig. 1Bi, Bii and Biii).

At the next stage, we investigated the strength of the camel and human promoters isolated from the constitutively expressed HSPA1L gene arranged in an inverted orientation in relation to the inducible members in the cluster (Fig. 1A). It was interesting to examine whether the HSEs located in the HSPA1A promoter somehow influenced the expression of the adjacent HSPA1L gene after HS. It has been previously shown that HSEs may in principle enhance transcription bidirectionally [23,27]. Therefore, we developed a series of seven constructs including four of that had the entire regulatory region between HSPA1A and HSPA1L genes differing by the length of the 5`-UTR (S3 Fig.). In the two other constructs, HSEs responsible for transcription induction after HS were experimentally deleted (S3 Fig.). Finally, in the last construct obtained from the human HSPA1L gene, a 112 bp sequence detected only in primates was deleted to monitor its possible impact on promoter activity (-142…-254 bp in human, see S1 Fig. and S3 Fig.). Luminescence measurements results demonstrated that the HSPA1L promoters of both species are not induced by HS in HEK293 and HDF cells, while the camel promoter exhibited at least 5-fold higher activity under normal conditions and after HS when compared with the orthologous human promoter (Fig. 1Biii). Interestingly, deletion of the 112 bp sequence in the human promoter led to a statistically significant increase (two-fold) in construct expression compared with the intact promoter (Fig. 1Biii).

To determine whether any transcription factors bind this sequence (112 bp) specific for the human promoter, EMSA experiments were performed using various labeled fragments of the HSPA1L regulatory region of both species with protein extract from HEK293 cells. Such analysis failed to demonstrate any binding of the human 112 bp sequence with regulatory proteins in HEK293 cells (Fig. 1C). We speculate that the observed enhanced luminescence after experimental deletion may be explained by the fact that a second Sp1-binding site becomes located closer to the transcription start after deletion of the 112 bp sequence (S3 Fig.).

Notably, partial deletion of the 5`-UTR decreases luminescence in the case of the human HSPA1A and HSPA1B and camel HSPA1B genes but has apparently no effect on camel HSPA1A genes and the constitutively expressed HSPA1L copies in both species (S3 Fig. and S4 Fig.).

Fig. 1B summarizes the data for the promoter strengths of the human and camel hspA1 gene cluster under normal conditions and after HS and demonstrates similar transcriptional activity for camel promoters in heterologous human cell systems, which suggest, among other explanations, a high level of conservatism and universality for HSPA1 group promoters in distant mammalian species.

S. singularior hsp70 promoters exhibit low activity in contrast with those of hsp83 in D. melanogaster cell culture

Previously, we have shown that all sequenced S. singularior hsp70 genes differ in their regulatory regions in contrast to the six known D. melanogaster hsp70 genes, which have nearly identical promoter sequences [21]. To reveal the role of the observed characteristic differences in the structure of the D. melanogaster and S. singularior hsp70 regulatory regions, we investigated the ability of these highly divergent species-specific hsp70 promoters to drive constitutive and heat-induced expression of the luciferase gene in Schneider-2 (S2) D. melanogaster cell culture.

To reach this goal, we developed constructs with the luciferase ORF under the control of hsp70 promoters with different architecture (S5 Fig.). Fortunately, the 3′-UTRs of all five hsp70 genes from S. singularior are unique, and in our previous study, we performed 3′-RACE-analysis and demonstrated that all five of the genes are transcribed under heat shock conditions [21]. In the construct designated hsp70S3, the luciferase ORF was placed under the control of the S. singularior hsp70S3 gene promoter. We also used a promoter from another S. singularior hsp70 gene (hsp70S4) to examine whether the low activity observed in D. melanogaster cells was due to a specific structure in the hsp70S3 gene or if it represented a characteristic of different Stratiomys hsp70 genes. Our analysis indicates that promoters of hsp70S3 and hsp70S4 genes exhibit similar “strengths” when tested in the luciferase reporter system. The reporter construct with a promoter from the D. melanogaster hsp70Aa gene served as a positive control in these experiments. With the exception of the upstream regulatory 5’-region, all constructs contained the 5’-UTRs necessary for normal translation under HS conditions (S5 Fig.). Previously, we demonstrated that the S. singularior and D. melanogaster Hsp70 mRNAs have similar stability [19].

Our experiments monitoring in vitro transcription levels (luminescence) demonstrated that the D. melanogaster hsp70Aa promoter has at least ten-fold higher activity in S2 cells, both under steady-state conditions and after HS (37°C) compared with the hsp70S3 and hsp70S4 promoters (Fig. 2) in the luciferase expression system. Because we detected drastic species-specific differences with regards to the hsp70 promoter activity in S2-based transient transcription assays, it was of significant interest to determine whether other hsp genes in these species follow the same pattern (i.e., much lower strength for S. singularior promoters). Therefore, we generated two constructs containing either the 318 bp sequence of the S. singularior hsp83S2 promoter (-155…+163 relative to the transcription start site) or the 297 bp sequence of the D. melanogaster hsp83 promoter (-147…+150). The constructs are depicted in S5 Fig., and the primers used are shown in S1 Table. Surprisingly, in contrast with the results of the hsp70 promoter studies, similar transfection experiments with subsequent luminescence measurements demonstrated that the hsp83 promoters of both species exhibit similar strength under normal conditions (25°C) and after HS (37°C) (Fig. 2). The obtained results coupled with recent data on the structure of Hsp83 promoters [22, 23] indicate that the hsp83 promoters of the compared distant Diptera species apparently preserve a high level of conservation; hence, they are capable of driving the transcription of the reporter gene with nearly equal efficacy in cells of the same host species (D. melanogaster).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. A. Various hsp70 promoter strengths based on the luciferase luminescence levels in S2 cells and the effects of experimental GAGA element insertion into the hsp70S3 promoter. B. The level of S. singularior and D. melanogaster hsp83 promoter activity in D. melanogaster S2 cells. *** p ≤ 0.001, ## p ≤ 0.01.

https://doi.org/10.1371/journal.pone.0115536.g002

S. singularior hsp70 promoters are capable of binding Drosophila HSF but not GAF in EMSA assay

It was interesting to reveal the molecular mechanisms underlying the dramatic differences in transcription activity between the D. melanogaster and S. singularior hsp70 promoters observed in vitro in Drosophila cell culture. First, it was necessary to determine whether the promoter of the hsp70S3 gene is capable of binding Drosophila HSF. This promoter contains three HSEs at positions similar to those in the D. melanogaster hsp70 genes but drastically different from the D. melanogaster hsp70 promoters in the context of surrounding sequences. Besides, hsp70S3 gene promoter has a different transcription initiator and a few substitutions in the HSEs (S2 Fig.) [21].

EMSA experiments with protein extracts from heat-shocked D. melanogaster flies and S2 cells are depicted in Fig. 3A. It is evident that sequences-94 to-39 bp relative to the transcription start sites in the S. singularior hsp70S3 and hsp70S4 promoters are capable of binding the D. melanogaster HSF factor, similarly to the orthologous region taken from the D. melanogaster hsp70 regulatory zone used as a positive control (Fig. 3A). Furthermore, the addition of anti-HSF antibodies prevents the HSF-HSE complex from migrating into the gel (supershift), verifying the binding specificity (Fig. 3A). Therefore, we conclude that the detected differences in promoter strengths apparently do not depend on the different abilities of the D. melanogaster and S. singularior hsp70 promoters to bind the Drosophila HSF.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. A. EMSA experiments with protein extracts from S2 cells and adult D. melanogaster flies exploring labeled fragments of the D. melanogaster hsp70Aa and S. singularior hsp70S3 and S4 genes. 1—Control (25°C), 2—heat shock (37°C), 3—heat shock + preimmune serum, 4—super shift with anti-HSF. The arrow indicates the position of the HSF-HSE complex. B. Recombinant D. melanogaster GAF protein effectively binds to the D. melanogaster hsp70 and S. singularior hsp70S3 promoters with the experimental insertion of three GAGA elements (lanes 1 and 4) but not with the wild-type hsp70S3 and hsp70S4 promoters (lanes 2 and 3). The arrow indicates the position of the GAF-GAGA complexes.

https://doi.org/10.1371/journal.pone.0115536.g003

GAGA-binding factor (GAF) is another candidate factor that may be responsible for the observed differences. GAGA elements are essential sequences found in D. melanogaster hsp70 promoters and a few other HS genes of various Drosophila species [24,28,29]. While GAF is necessary for the induction of hsp70 genes in Drosophila [24], canonical GAGA motifs were not detected in orthologous genes in S. singularior (S2 Fig.). On the other hand, we detected a few GAG or CTC sequences within the S. singularior hsp70 regulatory regions, and there was evidence that under certain conditions, such triplets may also bind GAF with low efficiency [30].

To determine whether Drosophila GAF is capable of binding S. singularior promoters, we performed EMSA experiments using purified recombinant D. melanogaster GAF. In this series of experiments, we used PCR fragments from the promoters of the hsp70S3 and hsp70S4 genes with orthologous D. melanogaster sequences as a positive control. The results demonstrated that the promoters of the hsp70S3 and hsp70S4 genes were not capable of binding D. melanogaster GAF, while, as expected, normal binding was observed for endogenous D. melanogaster promoter sequences containing functional GAGA elements and an hsp70S3 promoter with an experimentally inserted canonical GAGA site (Fig. 3B).

We hypothesized that the absence of functional GAGA elements in the S. singularior hsp70 promoter may contribute to the observed dramatic differences in the transcription of the hsp70 promoter-fused constructs under control conditions and after HS in D. melanogaster S2 cells.

Experimental insertion of GAGA elements into the S. singularior hsp70 promoter enhances its efficacy in the luciferase assay in vitro

To determine whether the presence of functional GAGA elements is the cause of drastic differences in the strength of the Drosophila and S. singularior hsp70 promoters in D. melanogaster cell culture after HS, we inserted different numbers of canonical D. melanogaster GAGA elements into the hsp70S3 regulatory region (see S5 Fig.) and monitored the transcriptional activity of the resulting promoters using a luciferase assay.

The insertion of a single functional GAGA-element in the plus orientation at position-65 bp from the transcription start site (similar to its localization in the D. melanogaster hsp70 promoters) did not significantly enhance hsp70S3 promoter activity after HS and only slightly increased promoter activity under control conditions (construct hsp70S3-GAGA+ in Fig. 2 and S5 Fig.). The insertion of a second GAGA element in the opposite orientation-145 bp from the transcription start site (hsp70S3-GAGA+/-) also did not influence the activity of the promoter after HS but resulted in a nearly four-fold increase in transcription under basal conditions. Furthermore, the construct (hsp70S3-GAGA+/-2) containing the insertion of two GAGA elements at-65 (“+” orientation) and-120 (opposite orientation) bp exhibited a two-fold increase in transcription after HS and a 10-fold increase under basal conditions in comparison with the original S. singularior hsp70S3 promoter in S2 cells. Therefore, the presence of two GAGA elements in the opposite orientation-55 bp from the transcription site resulted in transcription driven by the modified promoter under basal conditions, which is equal to that of the endogenous D. melanogaster promoter, and there was a mild increase in activity after HS (Fig. 2). Finally, the insertion of three functional GAGA elements at-65 (“+” orientation), -120 (“-”orientation) and-145 (“-”orientation) led to a 20-fold increase in transcription levels under steady-state conditions and a 5.6-fold increase after HS in comparison with the original hsp70S3 promoter (Fig. 2, hsp70S3-GAGA+/-3 construct). In other words, the hsp70S3 promoter containing the insertion of three functional GAGA elements similar to their localization in the D. melanogaster hsp70 promoter is capable of driving heat-induced expression in D. melanogaster S-2 cells with only two-fold less activity than the endogenous D. melanogaster promoter (Fig. 2). Characteristically, the modified hsp70S3 promoter in the S2 transient expression system significantly exceeded the activity of the wild-type endogenous D. melanogaster promoter under steady-state conditions. Importantly, the hsp70S3 promoter sequence containing experimentally inserted GAGA elements bound D. melanogaster GAF almost as efficiently as the D. melanogaster hsp70 promoter (Fig. 2). It was important to determine whether the pattern of the hsp70 activities observed in vitro in Drosophila cells (S2) using artificial plasmid-based constructs would coincide with the promoter activities determined in vivo in transgenic strains, which have normal nucleosome structure and complex chromatin architecture.

The difference in the TATA box structure of the S. singularior hsp70 promoter has no functional significance

Notably, the regulatory regions of all cloned S. singularior hsp70 genes are different from the D. melanogaster orthologs in the structure of the TATA box (TATATA in S. singularior vs TATAAA in D. melanogaster). To examine the possible functional significance of the observed differences in TATA boxes between the two species, a construct was generated in which the last T S. singularior hsp70 promoter was substituted by an A, making the motif identical to that in D. melanogaster. This construct also includes the insertion of three GAGA motifs and was designated hsp70S3-GAGA+/-3 (TATA) (S5 Fig.). Substitution of the hsp70S3 TATA box for that of D. melanogaster did not lead to increased activity for the hsp70S3 promoter containing three inserted GAGA elements (Fig. 2); hence, the observed difference in the TATA box structure most likely has no functional importance in this system.

S. singularior hsp70 promoter exhibits low activity in D. melanogaster transformed strains

Because the active state of chromatin and the nucleosome structure represent the most important factors providing the high activity of hsp70 gene promoters, it was important to determine whether the patterns of hsp70 promoter activity described in the above in vitro expression experiments exploring plasmid-based constructs resemble promoter strength in vivo.

To reach this goal, we compared the activity of endogenous hsp70 promoters with that of S. singularior in several transgenic strains of D. melanogaster obtained using P element mediated transformation as described in the Materials and Methods. Importantly, in all transformation experiments, we used strains with all copies of hsp70 genes deleted [31]. In the development of transgenic strains, we used the construct I(S) containing the ORF and 5’-UTR of the S. singularior hsp70S3 gene driven by its own intact promoter, the construct I(S-GAGA), which is similar to I(S) but with three GAGA insertions, designated the hsp70S3-GAGA+/-3 construct, and finally the construct II(M) in which the S. singularior hsp70 promoter and 5’-UTR were replaced by orthologous D. melanogaster sequences linked to the S. singularior ORF (Fig. 4). We have chosen the hsp70S3 gene for these comparative analyses because its promoter exhibits maximal deviation from the consensus structure of the D. melanogaster hsp70 promoter [21].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Left panel: General arrangement of constructs used in the transformation experiments.

The thick green arrows in I(S-GAGA) indicate the position of the inserted GAGA elements. HSE, GAGA elements and TATA boxes are indicated by square boxes of different color. Right panel: in situ hybridization of heat-shocked salivary gland chromosomes with the white gene fragment included in the constructs. The sites of the inserts are shown by arrows with puffs formed only in the strains (the bottom panel) containing constructs with D. melanogaster hsp70 promoters. In all panels, the heat shock puffs formed in the locations of major D. melanogaster hsp genes (i.e., 63B, 61C and 95D) that represent an internal control are indicated. 3B is the white locus that hybridizes with the labeled probe and represents an internal control for hybridization efficiency. In each strain, at least ten larvae were used for puff detection after HS.

https://doi.org/10.1371/journal.pone.0115536.g004

The hsp70S3 promoter includes three HSEs similarly spaced as in the D. melanogaster hsp70 regulatory region that contains four HSEs (S2 Fig.). However, in contrast with all D. melanogaster hsp70 promoters, the hsp70S3 promoter and other S. singularior hsp70 promoters do not contain canonical GAGA elements in the plus (GAGAGAG) or minus (CTCTCTC) orientation but includes only a few scattered triplets of the GAG or CTC type.

We obtained three transgenic strains with the S. singularior hsp70S3 ORF under its own promoter (I(S) transgene), designated Sa, Sb and Sc; two strains with the I(S-GAGA) transgene, designated 1G and 3G; and three strains with a construct in which the S. singularior promoter was replaced by a D. melanogaster variant (II(M) construct), designated Ma, Mb and Mc. The results of in situ hybridization with larval salivary gland polytene chromosomes from all three types of transgenic strains are depicted in Fig. 4 and Table 1. Notably, the obtained transgenic strains usually harbored one or two insertions. 5`-RACE analysis confirmed normal structure for the 5’-ends of mRNAs by sequencing resulting PCR fragments transcribed from the I(S), I(S-GAGA) and II(M) constructs in all obtained transgenic strains (data not shown). It has been demonstrated in many studies that after temperature elevation in the strains transformed by constructs under heat shock promoters typical heat shock puffs are formed not only at sites of well-known heat shock loci but in sites of the inserted constructs as well [7].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. List of transgenic strains used in the investigation.

https://doi.org/10.1371/journal.pone.0115536.t001

Heat shock experiments performed herein demonstrated that while normal prominent puffs are formed in D. melanogaster polytene chromosomes after HS (37°C) where the II(M) construct with an endogenous promoter was inserted, no puffs were observed at the sites of construct I(S) insertions where the Hsp70 ORF was under the control of the S. singularior wild-type promoter or at the sites of I(S-GAGA) transgene insertions (Fig. 4).

The extremely low heat-inducibility of the S. singularior hsp70 promoter in the D. melanogaster transgenic strains was further corroborated by Q-RT-PCR experiments. We detected low levels of hsp70 mRNA under normal conditions and a modest induction of its synthesis in the I(S) transgenic strains after temperature elevation (Fig. 5). Thus, in these strains, HS (37°C for 30 min) induced a slight induction in the hsp70 mRNA synthesis (3–4-fold). In Ma—Mc strains as expected, we detected several hundred fold induction after HS. Characteristically, similar patterns of induction in both types of transformed strains, i.e., I(S) and II(M), were detected in the larvae and in imagoes (Fig. 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. The results of RT-PCR experiments using RNA isolated from adult flies (A) or the third instar larvae (B) of D. melanogaster strains transformed with constructs containing different hsp70 promoters.

A significant dichotomy in the expression of the I(S-GAGA) construct in adults and larvae is evident, while all other constructs exhibit similar strengths at both stages. *** p ≤ 0.001.

https://doi.org/10.1371/journal.pone.0115536.g005

In contrast, while the strains transformed with constructs containing the S. singularior promoter with inserted GAGA motifs did not exhibit significant induction after HS at the imago stage, forty-fold induction was observed at the larval stage. However, the level of induction in the larvae of both strains transformed with the I(S-GAGA) constructs remained 6–7 times lower than in M-strains containing constructs with the endogenous D. melanogaster hsp70 promoter (Fig. 5B). Therefore, while in the case of in vitro transient expression, the insertion of three consensus GAGA elements into the S. singularior hsp70 promoter strongly enhanced its activity and rendered it comparable to the orthologous D. melanogaster promoter (Fig. 2) in in vivo experiments exploring transformed strains, a similar insertion resulted in modest induction observed only in the larvae (Fig. 5). Characteristically, this mild increase in promoter strength was not accompanied by heat-induced puff formation in contrast with the constructs of the II(M) type containing D. melanogaster endogenous promoter (Fig. 4). Thus, the inclusion of the consensus GAGA elements into the S. singularior hsp70 promoter does not lead to its normal functioning in the D. melanogaster genome in vivo but resulted in a modest increase in the promoter strength after HS observed only in the larvae of the transformed strains.

Sequence analysis

The programs NCBI BLAST and Vector NTI were used to align investigated sequences. The detection of regulatory motifs within the hsp70 and hsp83 promoters was performed with the MEGA 4.0 [46] and MatInspector (GENOMATIX Software) programs. Relevant sequence information concerning all investigated hsp70 genes including regulatory regions is accessible in GenBank. The GenBank accession numbers for hsp70S1, hsp70S2 and hsp70S3 are HQ184404 and HQ184405 for hsp70S4 and hsp70S5 [21]. The GenBank accession numbers for the two S. singularior hsp83 genes are JN627861 and JN627862, respectively [22]. The D. melanogaster hsp70Aa and hsp83 genes can be found in FlyBase in the context of the D. melanogaster genome. A cluster of camel HSPA1 genes has been deposited in GenBank under the accession number JF837187 [19], while human HSPA1 genes from the orthologous region are referenced in GenBank as BA000025.

Construction of luciferase reporter plasmids

Plasmids with fused promoter and luciferase reporter genes were constructed using a promoter/enhancer-free pGL4.10[luc2] basic vector (Promega). DNA fragments containing different portions of the hsp70 and hsp83 promoters were amplified using PCR with specific primers containing different restriction sites at the 5’-ends (S1 Table and S2 Table), and they were ligated upstream of the luciferase ORF into the pGL4.10 vector. DNA fragments previously isolated from λ phage genomic libraries of C. dromedarius and S. singularior [19,21] were used as a template for cloning the hsp promoters of these species. For PCR amplification of the D. melanogaster and human promoters, genomic DNA isolated from D. melanogaster flies and human HDF cell culture was used. Different overlapping primer pairs were used to obtain constructs with insertions of GAGA elements or substitutions within the TATA box of the S. singularior hsp70S3 gene upstream of the ATG in the camel HSPA1A gene (S1 Table). The promoters of the hsp83 genes were amplified by PCR from D. melanogaster genomic DNA and recombinant λ phage containing a part of S. singularior hsp83 cluster [22]. The resulting plasmids were transformed into competent E. coli DH5a strains for amplification. Endotoxin-free plasmid DNA was isolated using the Plasmid Plus Midi Kit (Qiagen). All constructs were confirmed by restriction analysis and DNA sequencing. Full schemes of all obtained constructs with indicated boundaries and directions are provided in S3 Fig., S4 Fig., and S5 Fig.

Cell lines, heat shock treatment and luciferase assay

For measurement of mammalian HSP70 promoter activity, reporter constructs were expressed in the human cell lines HEK293 and HDF [47,48] cultured in DMEM/F12 medium supplemented with 10% fetal bovine serum for luciferase reporter assay experiments. The activity of the Diptera hsp70 and hsp83 promoters was determined in D. melanogaster Schneider-2 cells (S2) in Schneider’s Insect Medium (Sigma) supplemented with 10% fetal bovine serum (HyClone). The TransFast (Promega) reagent was routinely used for the transfection experiments according to the manufacturer’s instructions. Both a firefly luciferase reporter gene construct (200 ng) and a pRL-SV40 Renilla luciferase construct (20 ng for normalization of transfection efficiency) were cotransfected into each well of 24-well plates. After transfection, human cells were incubated at 37°C for 48 h, placed in an incubator at either 37°C (control) or 43.5°C (heat shock) for 30 min and then incubated for 5 h at 37°C. S2 cells were incubated at 25°C and heat shocked at 37°C for 30 min with subsequent recovery for 5 h at 25°C. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) with The Reporter Microplate luminometer (Turner Designs). The ratio of Firefly to Renilla luciferase activities was taken as the activity of reporter construct. The activity of each reporter construct was measured at least in five independent experiments. The significance of differences in the reporter gene expression was estimated using Fisher’s test.

Drosophila fly strains and heat shock

For P element-mediated transformation experiments, we used the D. melanogaster (w[1118]; Df(3R)Hsp70A, Df(3R)Hsp70B) strain (strain 8841 in Bloomington Drosophila Stock Centre) in which all inducible hsp70 genes from the 87A and 87B loci were deleted [31]. Protein extracts for EMSA were obtained from the D. melanogaster w[1118] strain. Flies were reared at 25°C on standard sugar-yeast-agar medium containing propionic acid as a mold inhibitor. For heat shock, larvae or adult flies were transferred into preheated 50 ml plastic conical tubes (40 individuals/vial) for temperature stress in a water bath (37°C during 30 min) without anesthesia. After heat treatment, flies or larvae were frozen in liquid nitrogen immediately or after a recovery period and were used for RNA or protein extraction for 5’-RACE, Q-RT-PCR and EMSA experiments. For in situ hybridization, larvae were grown at 18°C on standard medium for 2 days before dissection, and for heat treatment, approximately 20 larvae were placed into 50 ml Falcon tubes with a moistened piece of filter paper and incubated in a water bath (37°C for 20 min).

The development of plasmids for transformation and construction of transgenic flies

The I(S) construct includes an EcoRV-XbaI fragment from the S. singularior hsp70S3 gene (4663 bp), containing an ORF (1917 bp), a 5’-UTR and 3’-UTR (230 and 247 bp correspondingly), a 5’-flanking region (2117 bp) and a 3’-flanking (152 bp) sequence. This fragment was obtained from λ phage containing a portion of the S. singularior hsp70 cluster [21], cloned into the pBluescriptSK+ vector and then subcloned into pCaSpeR5 transformation vector. To obtain an I(S-GAGA) construct with three GAGA element insertions, step-by-step PCR reactions with overlapping GAGA-containing primers (see S1 Table) were performed. Resulting DNA fragments were inserted into the I(S) construct with replacement of the wild-type promoter.

In construct II(M), we substituted the S. singularior hsp70 promoter and 5’-UTR with a 1557 bp PCR fragment containing the promoter and 5’-UTR region of the D. melanogaster hsp70Aa gene (primer sequences are given in S3 Table). Next, we excised a PstI-XbaI fragment containing the hsp70S3 ORF with a 3’-UTR and 3’-flanking sequence from the I(S) construct, ligated this fragment with the BamHI-PstI excision of the D. melanogaster PCR fragment into pBluescriptSK+, and finally subcloned the resulting sequence into pCaSpeR5. All constructs were confirmed by DNA sequencing.

Expression and purification of the recombinant D. melanogaster GAF protein

The histidine-tagged GAF fusion protein (His-GAF) contained amino acids 5–519 of D. melanogaster GAF fused in frame with six histidine residues from the pET-28a vector (Novagen, USA). The construct was kindly provided by Dr. E.M. Baricheva of the Institute of Cytology and Genetics, Russia, Novosibirsk. The Escherichia coli BL21 strain was used for protein expression. His-GAGA protein purification is described in [25].

Preparation of protein extracts

Protein extracts were obtained from control and heat-shocked adult D. melanogaster flies, S2 cells and HEK293 cells. Up to 10⁸ S2 or HEK293 cells were collected after heat shock or from control conditions (see above) and frozen in liquid nitrogen and further processed as described in [49]. D. melanogaster flies were taken from control conditions or stressed at 37°C and processed as in [25].

Electrophoretic mobility shift assay (EMSA)

Protein extracts from S2 and HEK293 cells were used in EMSA experiments with fragments from hsp70 promoters from Diptera and mammalian species (camel, human), respectively. EMSA with protein extracts from D. melanogaster flies and from S2 cells (10 μg of total protein) was performed as in [25] with ³²P-labeled oligonucleotides probes of hsp70S3, hsp70S4 and hsp70Aa sequences (see S3 Table for primer sequences). Drosophila anti-HSF rabbit polyclonal antibodies (kind gift of Prof. C. Wu, NIH) were used in supershift experiments. HEK293 protein extracts (10 μg) were shifted with ³²P-labelled PCR fragments from human or camel HSPA1L promoters (see S3 Table for primers sequences). For EMSA with recombinant D. melanogaster GAF, oligonucleotides corresponding to the S. singularior hsp70S3 and hsp70S4 and a D. melanogaster hsp70Aa promoter were amplified by PCR with specific primers (S3 Table) in the presence of α³²P-dATP. The binding reaction mixture contained 1 μg of recombinant GAF, 5 ng of (³²P)-labeled DNA probe, and 10 ng of poly (dI-dC) in the band shift buffer [25]. The reaction mixtures were separated in a 4.5% AA/Tris/borate/EDTA gel. Gels were dried immediately after electrophoresis and exposed to X-ray film.

P Element-Mediated Transformation of D. melanogaster

Plasmid DNA of the constructs I(S), I(S-GAGA) and II(M) was purified using the QIAprep Spin Miniprep Kit (Qiagen) and used for embryo injection as previously described [50]. Transposase activity was provided by the helper plasmid Turbo Δ2–3, containing a stable genomic source of P element transposase from D. melanogaster [51]; the recipient embryos were from D. melanogaster strain 8841, which lacks all hsp70 genes [31]. Adults emerging from the injected embryos were back crossed with flies from the 8841 strain of the opposite sex, and the eye color of their progeny was examined. Transformed strains homozygous for the transgene were established using full sibling mating. The localization of homozygous transgene copies and ability of construction to puff generation was performed using in situ hybridization with a white biotinylated probe as described in [52]. Transgene intactness was checked by PCR with primers specific for the P element and hsp70 flanking sequences, and transcription start sites of all three constructs in transformed strains were determined by 5’-RACE as described in [21]. For primer sequences, see S4 Table.

Quantitative Real-time PCR

Total RNA was isolated from larvae or adult flies using TRIzol (Invitrogen). cDNA was prepared from 1 μg of total RNA treated with DNase 1 (Ambion, USA) using random decamer primers and MMLV reverse transcriptase (Evrogen). Experiments were performed using an ABI PRISM7500 Sequence Detection System (Applied Biosystems, USA). Each reaction was performed with a minimum of 3 biological replicates and performed using 2.5 X SYBR Green 1 PCR Master Mix in the presence of ROX reference dye (Syntol) in accordance with the manufacturer’s protocol. Reaction efficiency was evaluated and was at least 95%. Quantification was normalized to the ubiquitously expressed rpl32 (ribosomal protein L32) and ef1a (elongation factor 1α) genes, and the ddCt method was used to calculate relative expression levels. Primer sequences are provided in S3 Table. The significance of differences detected in the experiments was estimated using Fisher’s test.